Author: tenderloin

Time for China to build an EUV Lithography system for sub 7nm chips [Copy link] 中文

Rank: 8Rank: 8

Post time 2020-1-16 08:03:42 |Display all floors
ASML started small as a part offshoot from Phillips which was not just into light bulbs but also x-ray machines. EUV is just wavelength 13.5 nm UV which is close to x-ray so what ASML has done with its EUV lithography was to tap back to the roots of its cofounder.
Where ASML has expanded since its formation within the last four decades was by making strategic acquisitions of or collaborations with Carl Zeiss, Brion, Hermes Microvision, Mapper and Cymer which were essential for its lithographic expertise in precision mechatronics, something I had mentioned together with actuators and sensors as very important.. many years ago, in this very forum.

By so doing, ASML has carved its niche in microchip making machines by developing expertise in overlaying, double patterning and precision production software writing as it moves to push the limits of Moore's Law.

If a small Dutch outfit like ASML could specialise so much in such a short time with so few just as the British ARM had similarly done with its software, I can't see why China can't do more and faster.  ASML's inflated R&D budget of £2 billion a year is peanuts compared to the billions of Yuan which China has invested to home-grow her own semiconductor industry.. but she must however be extra careful not to throw good money on bad enterprises with weak ideas... because this vision is today of critical importance to the future of China, namely technology supply chain autonomy to blunt the US' determination to finish off China by cutting off its suppliers and those of other countries from China's IT and mobile (thus payments) industries, hence national economy, thus international and financial sovereignty.

So today it is still sand, light and software. Zeiss is used in Huawei which is trying new things. There must be better coordination, more focused use of human brains and funds, and more strategic pipelining of real expertise in the road ahead to develop one's own chip industry from end to end, not just throw money and incentives around... because time is the factor and time changes the goals to achieve from one year to the next. The US for example is now about to subsidise heavily into 5G while using every dirty trick it has to thwart other countries from accepting Huawei.

If China wants to build her own EDA (cad/cam AI rebooted) expertise in order to build her own synopsys, cadence design, mentor graphics, ansys, qualcomm, micron, intel, even foxconn, she has think out of the box about the whole industry and what it's future can be in a technologically decoupled world where there will be disintegrated, discrete and different technologies set to different designs creating a technology disparity where advanced countries will buy few, expensive and high-performing devices while the rest of the world will buy cheaper, reliable, mass-made and nimbly scalable equivalents.

Just to throw a thought.. what if it's not sand and light but still open-source but not RISC?


Use magic tools Report

You can't reply post until you log in Log in | register

BACK TO THE TOP
Contact us:Tel: (86)010-84883548, Email: blog@chinadaily.com.cn
Blog announcement:| We reserve the right, and you authorize us, to use content, including words, photos and videos, which you provide to our blog
platform, for non-profit purposes on China Daily media, comprising newspaper, website, iPad and other social media accounts.